Luminos · Data and AI consultancy
We engineer the pipelines, build the analysis and ship the dashboards, machine learning and AI systems your team relies on every day.
Partnering with UN Sri Lanka and WFP Sri Lanka
We partner with UN Sri Lanka and WFP Sri Lanka on the data systems behind their programme reporting.
We work with WFP Sri Lanka on field data pipelines and monitoring dashboards, so programme teams can follow activities by district as data arrives.
Sample data for demonstration. Select a district or change the measure.
Each service stands on its own, but they work best together: clean pipelines feed reliable analysis, which feeds dashboards, models and AI you can trust.
We design Power BI and web dashboards around the handful of numbers your team decides on, then wire them to live data so nobody rebuilds a report by hand again.
An interactive climate dashboard built on NOAA's Oceanic Niño Index. It tracks how warm the tropical Pacific is today, compares this year with the strongest El Niño years on record, and ranks every episode since 1950.
Overview: 75 years of the Oceanic Niño Index
Four working examples of what we build. Change the filters, click the bars and hover the charts. All figures are sample data.
| Period | Seasons | Peak | Strength | |
|---|---|---|---|---|
| SON 2014 – AMJ 2016 | 20 | +2.59 °C | Very strong | |
| AMJ 1997 – MAM 1998 | 12 | +2.37 °C | Very strong | |
| AMJ 1982 – MJJ 1983 | 14 | +2.14 °C | Very strong | |
| MJJ 2023 – MAM 2024 | 11 | +1.99 °C | Strong | |
| MJJ 1972 – FMA 1973 | 10 | +1.84 °C | Strong | |
| MJJ 1965 – FMA 1966 | 10 | +1.82 °C | Strong | |
| MAM 1957 – AMJ 1958 | 14 | +1.69 °C | Strong | |
| MJJ 1991 – MJJ 1992 | 13 | +1.54 °C | Strong | |
| JJA 2009 – FMA 2010 | 9 | +1.50 °C | Strong | |
| ASO 1986 – DJF 1988 | 17 | +1.49 °C | Moderate | |
| JJA 1963 – JFM 1964 | 8 | +1.19 °C | Moderate | |
| JJA 2002 – JFM 2003 | 8 | +1.19 °C | Moderate | |
| ASO 2018 – MJJ 2019 | 10 | +1.05 °C | Moderate | |
| MJJ 1951 – DJF 1952 | 8 | +1.03 °C | Moderate | |
| SON 1968 – AMJ 1969 | 8 | +0.98 °C | Weak | |
| SON 1994 – JFM 1995 | 5 | +0.97 °C | Weak | |
| ASO 2006 – DJF 2007 | 5 | +0.88 °C | Weak | |
| ASO 1976 – JFM 1977 | 6 | +0.84 °C | Weak | |
| SON 1977 – JFM 1978 | 5 | +0.84 °C | Weak | |
| MAM 1953 – NDJ 1953 | 9 | +0.78 °C | Weak | |
| SON 2019 – FMA 2020 | 6 | +0.75 °C | Weak | |
| JAS 2004 – JFM 2005 | 7 | +0.71 °C | Weak |
Peak ONI: weak 0.5–0.9, moderate 1.0–1.4, strong 1.5–1.9, very strong 2.0 °C and above.
Source: NOAA Climate Prediction Center, Oceanic Niño Index (3-month running mean of ERSST.v5 sea surface temperature anomalies in the Niño 3.4 region), 919 seasons from DJF 1950 to JJA 2026. Episodes follow the historical rule of at least five overlapping seasons beyond ±0.5 °C. Since February 2026 NOAA uses the Relative Oceanic Niño Index (RONI) for official ENSO status, so official labels can differ from this view.
Websites and platforms for public bodies and businesses in Sri Lanka and the UK. Pick a card to see the project.
Drag the handle to compare a typical monthly report before and after we automate it.
Every night, data moves from your systems to your dashboards. Run it here, then try it with bad data to see how problems get caught before anyone sees them.
Press “Run pipeline” to start.
Select a layer to see what it involves.
Most dashboards and models fail because of what sits underneath them. We work across the whole stack, so the insight at the top is only as confident as the data allows.
Every engagement starts from the decisions you need to make, not from a tool we want to sell.
Code lives in your repositories, models are documented, and your team can run it without us.
If the data cannot support the answer yet, we tell you early and fix the foundation first.
Move the sliders to match your team. Automated pipelines and dashboards usually remove most of the copying, cleaning and formatting behind recurring reports.
A rough estimate based on your inputs, not a quote.
Discuss your reportingFour stages, with something working to review at the end of each.
We map your data sources, the decisions you need to make and what is blocking them today.
You get a written plan: architecture, scope, timeline and what done looks like.
We deliver in two-week cycles and show working results at the end of each one.
Documentation, training and a support window so your team owns what we built.
Yes. Most of our work runs remotely, with regular video check-ins and shared project boards, so location is rarely a constraint.
Yes. We often start by fixing the data model and refresh setup behind existing reports before building anything new.
We work inside your cloud accounts wherever possible, follow your access policies, and keep personal data out of development environments.
A focused dashboard or pipeline usually takes 3 to 6 weeks. Machine learning and AI projects start with a 2-week feasibility phase so you know what is realistic before committing further.
Yes. We offer monthly support plans for monitoring, fixes and small improvements, or we can train your team to take it on fully.
Tell us the decision you need to make and where your data lives today. Email us or give us a call, and we'll reply within two working days.
